W3C home > Mailing lists > Public > public-xg-htmlspeech@w3.org > November 2011

Another translate example

From: Michael Bodell <mbodell@microsoft.com>
Date: Thu, 3 Nov 2011 07:04:35 +0000
To: "public-xg-htmlspeech@w3.org" <public-xg-htmlspeech@w3.org>
Message-ID: <22CD592CCD76414085591204EB19F4E8239D1397@TK5EX14MBXC262.redmond.corp.microsoft.com>
I worked through the translate example in the HTML as well.  The thing I didn't do, and haven't seen any of the examples do yet, is pull the specified HTML design through to the protocol.  I.e., what are the C->S and S->C messages that get sent with this example and how does the API react to these messages.



for instance:



S->C: web-speech/1.0 END-OF-SPEECH 2343 IN-PROGRESS

Resource-ID: recognizer

Source-time: 2011-11-3T16:01:23.456+1:30



maps to the

speechend event being raised with data equal to the *UA* mapping of the recognizer time



In this example that results in the progress bar being set to 60.



or when the sir.start() is executed the following interaction happens:



C->S: GET /reco HTTP/1.1

Host: example.org

Upgrade: websocket

Sec-WebSocket-Protocol: web-speech/1.0



...







I fear that if we don't trace an example through the protocol and back we'll be missing the connections between the two currently disparate examples.



Also, we might not catch all the missing places.  For instance, the js api right now doesn't really specify anything about timing information in speechend, it is quite underspecified really.



Received on Thursday, 3 November 2011 21:43:41 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:16:51 UTC